​

Boost Sales and Efficiency : How to Improve Forecast Accuracy for your Cannabis Business

Hey Reader,

Did you know that improving forecast accuracy by just 2% can reduce inventory costs by up to 1%? That’s a powerful supply chain rule of thumb that holds true across industries. Better forecasts lead to less overstock, fewer out-of-stocks and ultimately, increased sales and customer loyalty.

Many dispensary operators we talk with use rolling daily or weekly averages to predict sales and demand. While this method is quick and simple, it comes with a major drawback—forecast errors can range from 17% to 82%, depending on the timeframe and the level of forecasting. Clearly there are opportunities for improvement, and a growing urgency as our industry continues to mature and focus more attention on costs and supply chain efficiencies.

How do you know if your forecast is accurate? The key is to measure the Mean Average Percent Error (MAPE) which is the percentage difference between your forecast and actual values. For more about MAPE I encourage you to visit this link.

Some in our industry don't need forecasting yet. A grower in New Jersey told me he won’t need to forecast demand for the next two years because he expects every crop to sell out. For the rest of us, achieving high levels of forecast accuracy is hard work. Here’s why:

• Limited data: Advanced models require at least 2 years of historical sales data, which many newer dispensaries lack. Even when the data is available, it is often so big that it can be difficult to manage.
• Events and holidays: Industry events like 420 and Green Wednesday create huge sales spikes that are growing more important each year, as can be seen in the chart below:

• Limited-run and new products: These disrupt sales data and trends, making forecasting tricky and requiring estimations or imputed data.
• Supply chain issues: Out-of-stocks distort demand signals, especially at the SKU level, and must be factored in.
• Choosing the right model: There are dozens of proven specialized forecasting models. Finding the right one(s) and optimizing them for your data and business can be a challenging iterative process.

At Palomar, we’ve developed a cannabis-specific forecasting methodology that addresses these challenges by considering in-stock rates, events and holidays, promotions and price sensitivity, new & rotating products, seasonality and trends. By maximizing accuracy at every level and considering data quality and availability, we help clients to improve accuracy by 10% to 55%+ compared to their current forecasts. Here are some examples from a recent project:

• category level : +27% improvement to 88% forecast accuracy
• sub-category level : +23% improvement to 80-83% forecast accuracy
• brand level : +55% improvement to 70-82% forecast accuracy

As the cannabis industry matures, accurate forecasting is becoming a must-have tool for operators to find efficiencies, reduce costs and improve margins. Whether you’re just starting to explore or looking to fine-tune your current approach, we’re here to help.

Get in touch with us to discuss your forecasting needs. Feel free to reach out to me directly at nick@palomargroup.ai.

Be well!

Nick

​​​

Palomar Group, NW Washington DC 20008 United States
​Unsubscribe · Preferences​